Genetic evidence for a role of hexokinase isozyme PII in carbon catabolite repression in Saccharomyces cerevisiae.

A mutant of Saccharomyces cerevisiae that was selected for resistance to carbon catabolite repression also had reduced hexokinase activity. Hexokinase isoenzymes were purified from mutant and wild type cells. The specific glucokinase and hexokinase isozyme PI were present at normal levels in mutant and wild type, but no hexokinase isozyme PII activity was detected in the mutant. Staining for enzyme activity after electrophoresis of crude extracts also indicated that hexokinase PII was absent in the mutant. Mutant and wild type segregants gained by tetrad analysis were investigated electrophoretically. Staining for enzyme activity confirmed that catalytically inactive hexokinase PII and the defect in carbon catabolite repression always co-segregated. The results support the hypothesis that hexokinase PII might mediate carbon catabolite repression.

In the yeast Saccharomyces cerevisiae enormous differences in the activities of certain enzymes have been reported between cells growing on glucose and on nonfermentable carbon sources, such as ethanol or acetate. Such differences were observed for some enzymes of the tricarboxylic acid cycle and glyoxylate cycle , respiratory enzymes , gluconeogenic enzymes (Gancedo et al., 1965;Witt et al., 1965;Gancedo and Schwerzmann, 1976), cy-glucosidases (Wijk and van Ouwehand, 1969), and invertase (P-fructofuranosidase) (Gascon et al., 1968). This phenomenon has been called carbon catabolite repression in analogy to the situation defined in bacteria by Magasanik ( 1961).
A very effective selection system for isolating mutants resistant to carbon catabolite repression was described by Zimmermann and Scheel (1977). Three mutant classes were identified (Zimmerman and Scheel, 1977;Entian and Zimmermann, 1980). (a) hex2 mutants were no longer repressible by glucose for invertase, maltase, malate dehydrogenase, and respiratory enzymes. Hexokinase activity was strongly decreased (Entian et al., 1977). ( b ) hex2 mutants had a similar defect in carbon catabolite repression, but their hexokinase activity was strongly elevated (Entian and Zimmerman, 1980). Elevation of hexokinase activity could be attributed to increased hexokinase PI1 synthesis (Entian, 1981). Additionally, * This work was supported by Deutsche Forschungsgemeinschaft and by Fonds der Chemischen Industrie. The costs of publication of this article were defrayed in part by the payment of page charges.
This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Dedicated to Professor H. Holzer on the occasion of his 60th birthday in acknowledgment of his fundamental research on the regulation of yeast metabolism. the maltose uptake system was disregulated in hex2 mutants, which led to a strong inhibition of cell metabolism by maltose . (c) The third class of mutants, called cat80, was defective in repression of invertase, maltase, and malate dehydrogenase, but the hexokinase activity was like wild type (Entian and Zimmermann, 1980).
According to the hypothesis of Magasanik (1961), carbon catabolite repression is triggered by an accumulation of catabolic derivatives of hexoses. However, determination of such metabolites during growth on hexoses in these catabolite repression-resistant mutants did not indicate such a simple situation (Entian and Zimmerman, 1980). The abnormal hexokinase activities in hexl and hex2 mutants suggested an important role of hexokinase in carbon catabolite repression.
There are three enzymes that can phosphorylate D-glucose in S. cerevisiae. One is specific for glucose and is called glucokinase (Maitra, 1970;Ramel et al., 1971). The others, called hexokinase isozymes PI and PII, are nonspecific and can phosphorylate glucose, fructose, and mannose. PI and PI1 can be distinguished by their different ratios of fructose over glucose phosphorylation. Q F/G' for PI is about 3 and for PI1 is about 1.2 (for a review, see Colowick, 1973  The abbreviations used are: Q F/G, ratio of fructose over glucose phosphorylating activity; PMSF, phenylmethylsulfonyl fluoride; MTT, 3-(4,5-dimethylthiazoyl-Z)-2,5-diphenyl tetrazoliumbromide; EDTA ethylenediaminetetraacetate.

Enzyme and Protem Determlnatlon
Hexoklnase activlty with glucose as substrate was tested according to Bergmeyer 119701. When fructose was used a s Maltase and Invertase were assayed according to Zlmmermann a substrate, 0.5 U/ml phosphoglucoseisomerase were added. et al. (1977). Specrfic activity is defined a s umol substrate converted per mln (units) per mg proteln. Proteln was determmed as described by Lowry et al. (19511 or durlng chromatography by direct measurement of absorbance at 280 nm.

Cell Dlsruptlon
K-Po4 buffer. 120 ml suspension and 240 ml glass beads 220 g of cells (wet weight) were suspended in 100 ml (drameter 0 . 5 mm, Braun-Melsungen) were hornogenlzed at hlgh speed in a laboratory blender (Waring 1 BaW-GI for 1 0 min. The temperature did not exceed 25'12. Frozen pleces of K-Po4 buffer were added to the suspensron for coollng. Cell breakage was about 70% as determined rnlcroscoplcally. The supernatant was poured off, and the glass beads washed twice wlth 100 m l K-POd buffer to regain adsorbed proteins. All supernatants were mixed and centrlfuged (10 min, 24 000 g , Sorvall RC5). The supernatant was diluted to 1.2 1 wlth x-PO4 buffer and used as crude extract.
was opened. The flow rate was adlusted to 200 ml/h. When When a sedlment of about 0 . 5 cm had settled, the outlet the suspension had sedlmented (bed height 7.5 cm) a retrograde ammonium sulfate gradient (75% to 35%. total volume 5 1, fractlon volume 16 ml) was passed over the column. Each collection tube contained 1.25 ml K-PO4 buffer to avoid precipitatlon caused by evaporation. Actlve fractions were pooled and had about 5 5 % saturation. Saturatlon was readjusted to 80%. After centrlfugatlon ( 3 0 min, 27 000 g . Sorvall RC5l the sediment was dissolved In 125 ml succinate buffer and desalted on a Sephadex G 25 column (Pharmacla, diameter 9 cm, bed height 17.5 cm).

DEAE Sephacel Chromatography
The enzyme solutlon was passed over a DEAE sephacel ion exchanger (Pharmacla, diameter 2 cm, bed hexght 20 cml, equilibrated against succinate buffer and eluted by a linear salt gradient (total volume 400 ml, 0-0.2 M KC1). This high capacity exchanger (1.4 meq/g) gave good purification at high flow rate (225 ml/h). Fractions of 2 ml were collected.
in the Ultra Phor apparatus (Colora, Lorch. F.R.G.) Gels were stained for protein wlth Coomassie blue.
Hexokinase activity was detected by the formazane procedure IThorne et al., 1963). A method developed by Frohlich and Entian (unpublished results) was used.

Purification of Hexokinases
To lnhlbit proteases (Easterby and Rosemeyer, 19721 a 40 mM solution of PMSF ( p h e n y l m e t h y l s u f l o n y l f l u o r l d e l (Sigma) in absolute ethanol was slowly added to glve a final concentration of 2 mM. Nucleic acids were precipitated by streptomycm sulfate (Serva, Heidelberg) at a final concentratlon of 1%. After 20 min stlrrlng, the suspension was centrlfuged 1 3 0 mln, 27 000 g).

DE 52 Chromatography
Active fractlons were pooled and diluted l:l0 t o allow direct additlon onto a DE 52 column (Whatman, 6larneter 2 cm, bed helght 15 cml. In contrast to Barnard (19751. the DEAE cellulose was equilibrated with succinate buffer, pH 6 . 3 , because hexokinase P I did not blnd satlsfactorlly at pH 5.8. when the diluted pool was used. A slightly concave pH gradient was used for elutron (200 m l pH 6.3 -200 ml pH 4.6, flow rate 3 0 ml/h). Actlve fractlons of and purifled simultaneously. The pools were carefully Isoenzymes PI and PI1 (fractlon volume: 2 m l l were pooled another DE 52 column(d1ameter 1 cm, bed hexght 10 cm, tltrated to pH 6.3 uslng conc. NaOH and passed onto equlllbrated with succinate buffer). A 200 m l salt gradient (0-0.2 M KC1) was used for elutlon (fractlon volume 1 ml, flow rate 20 ml/h). Active fractrons were pooled, diluted 1:lO and finally passed onto a s~m l l a r column. Hexoklnase activlty was eluted by a pH-gradient (200 ml. pH 6 . 3 -4.6, fractlon volume 1 ml, flow rate 20 ml/hl. Profiles of protein and hexokinase actlvlty were identical ( Fig.5 and 6, see also Table 3 ) .

Comparative Purification of Hexokinases from an hexl Mutant and Wild Type Cells
Ammonium Sulfate Chromatography-For ammonium sulfate chromatography (King, 1972), proteins, which had been salted out by high ammonium sulfate concentrations, were adsorbed to Celite and packed into a column (see under ''Experimental Procedures"). The proteins were eluted by a retrograde ammonium sulfate gradient. Glucokinase, which precipitates below 50% ammonium sulfate saturation, and hexokinases, which precipitate above 50% ammonium sulfate, were completely separated. As shown in Fig. la, the hexokinase isozymes PI and PI1 were also partly separated in extracts of wild type cells. Glucokinase was eluted at 46% saturation, hexokinase PI at 60% saturation, and hexokinase PI1 at about 53% saturation. In the her1 mutant (Fig. lb), glucokinase was eluted at 43% saturation, similar to results with the wild type extracts. However, the elution profile of the hexokinases differed from wild type. Q F/G ratios and ammonium sulfate concentration of resolution indicated that only isozyme PI was present in the hexl mutant.
For further purification the hexokinase PI and PI1 fractions were pooled from the wild type and the hex1 mutant, respectively. These pools were placed on a DEAE-Sephacel ion exchanger and eluted with a linear salt gradient (0-0.2 M KCl). This fast chromatography did not separate the isoenzymes, but yielded a good purification (see Table 111).
Separation of Hexokinases Isozymes PI and PII-Wild type isoenzymes PI and PI1 were separated by elution with a pH gradient on DEAE-cellulose according to Barnard, 1975, with modifications (see under "Experimental Procedures").
PI was eluted at pH 5.5 and PI1 at pH 5.1 (Fig. 2a). In the hexl mutant, only isoenzyme PI was eluted at pH 5.5 (Fig.  2b). No further activity was resolved, even after the addition of 1 M KC1 to elution buffer. This indicated that no additional activity was retained on the column. Recovery of enzyme activity was 93% in the hexl mutant. These results clearly indicated that isoenzyme PI1 was not present in the hex1 mutant.

873
The separated isoenzymes were further purified using two smaller DEAE-cellulose columns, which were eluted with a salt gradient and a pH gradient, respectively. In the hexl mutant and wild type preparations, isoenzyme PI had the same elution behavior. No additional enzyme was detected in hexl mutant. Isoenzyme PI1 from wild type cells was purified similarly. These purification procedures yielded single bands for hexokinases PI and PI1 on polyacrylamide disc electrophoresis (Fig. 3).
Depending on the purification conditions, hexokinases PI and PI1 can change their electrophoretic behavior. Proteases in crude extracts, which can modify PI and PI1 without loss of catalytic activity, are responsible for these changes (see Colowick, 1973). Only direct comparison by electrophoresis between crude extract enzymes and purified enzymes could detect those proteolytic modifications during enzyme preparation. Direct staining of gels for hexokinase activity was unsatisfactory, especially with fructose as substrate, because the two indicator enzymes needed caused extensive diffusion bands. Instead, the gels were cut into 1 mm slices, and the enzyme activity was assayed in each slice. No differences in electrophoretic mobility between crude extract enzymes and purified enzymes PI and PI1 were detected in wild type cells. The positions of hexokinase activity and the positions of protein bands after purification were identical (Fig. 4, a, b, and c). hexl crude extracts had only PI activity. Purified hexokinase PI from the hexl mutant and the PI of mutant crude extracts had the same electrophoretic mobility as hexokinase PI of wild type (Fig. 4, b, d, and e). This also indicated that hexokinase PII, if present at all, was not catalytically active in the hexl mutant.

Co-segregation of the Catabolite Repression Defect and Loss of PII Activity in hexl Mutants
After an appropriate cross, diploid cells were obtained that contained mutant allele hexl and wild type allele HEXl. After sporulation and tetrad analysis, segregation of mutant allele hexl was followed. All of the segregants of three tetrads were investigated by electrophoresis and staining for hexokinase activity. All of the segregants that were defective in carbon catabolite repression also had no hexokinase PI1 activity (Table I), which confirms that a defect in carbon catabolite repression and the loss of hexokinase PI1 activity are directly associated.

Substrate Saturation Kinetics of Hexokinase PI from the hexl Mutant and Wild Type Cells
As shown in Table 11, K , values for glucose, fructose, and ATP as substrates were the same for hexokinase PI from the hexl mutant and the wild type cells. This gave further evidence that decreased hexokinase activity in the hexl mutant was not due to alterations in the properties of hexokinase PI.

DISCUSSION
The results described here show that the decreased hexokinase activity in hexl mutants is attributable to an absent or catalytically inactive hexokinase PII. The activity and kinetic behavior of hexokinase PI was not affected. About 500 spontaneous mutants, which are allelic to the hexl mutant have so far been isolated according to the very efficient selection system of Zimmermann and Scheel(l977). At least 50 of them were also tested for hexokinase activity. In all cases hexokinase activity was additionally reduced. Since hexl mutant was isolated, about 100 tetrads have been investigated. In all cases low hexokinase activity and defect in carbon catabolite repression co-segregated. All efforts to identify alterations in the concentrations of glycolytic derivatives of glucose in carbon catabolite repression mutants have been unsuccessful (Entian et al., 1977;Entian and Zimmermann, 1980). This made the hypothesis of Magasanik (1961) that carbon catabolite repression might be triggered by an accumulation of glycolytic derivatives of hexoses unlikely. On the other hand, the properties of the hexl mutant and the central role of hexokinase in glycolysis give evidence that hexokinase PI1 might trigger carbon catabolite repression. We suggest that hexokinase PI1 is a bifunctional enzyme, having a catalytic and a regulatory function. As proposed by Zimme~-mann,~ we propose a role for hexokinase PI1 as the "recognition site" of carbon catabolite repression. Such a role would require that the enzyme also be able to respond in some way to the availability of hexoses or their derivatives.
Indeed, this proposed regulatory function might be associated with some known effects of hexoses and glycolytic intermediates on hexokinase PI1 (for a review, see Colowick, 1973). The addition of glucose leads to dissociation of the enzyme F. K. Zimmermann, personal communication.
The presence of ATP induces association of hexokinase PI1 (Shill et aL, 1974). X-ray structure analysis by Steitz et al. (1977) gave evidence for an allosteric activator site for ATP which lies between the subunits. At low ATP concentrations a pronounced activation by 3-phosphoglycerate and phosphate has been reported (Kosow and Rose, 1971). This, however, was probably an artifact resulting from chelation of inhibitory aluminium present in ATP preparations (Viola et al., 1980). Furthermore, two conformations of hexokinase PI1 have been described (see Barnard, 1975). Possibly one of these properties of hexokinase PI1 might be involved in the sequence of events that bring about carbon catabolite repression.